vriforecastingThis report provides the forecast of 1-week ahead confirmed cases for
the 4 prevalent respiratory viral cases in BC (Flu-A, Flu-B, RSV and
SARS-CoV-2). Forecast reports were generated using both the
weekly-aggregated data population (PLOVER) and the weekly-aggregated
pediatric only population data (PHRDW). vriforecasting uses
EpiEstim as the backend to forecast weekly confirmed cases,
which forecasts cases by estimating the instantaneous reproduction
number using the Cori
method and then simulating the number of new cases using fixed
exponential growth over a given time period.. The report is divided into
two sections: 1-week forecast for the current season of respiratory
cases as well as validation plots and metrics for the performance of the
forecasting model on historical data to help in the assessment of model
performance.
This section presents the 1-week ahead model forecast for the current season for both PHRDW and PLOVER data. The 1-week ahead forecasted confirmed cases is reported with the 95% and 50% prediction interval. The last estimated reproduction number is also reported with the 95% confidence interval.
The 1-week ahead forecast value of confirmed cases for 2024-04-28 is: 115 cases/week
The 95 % prediction interval for this forecast is (76-162)
The 50 % prediction interval for this forecast is (100-130)
The last estimated Rt value with the 95% confidence interval is: 1.15 (1.07-1.23)
The 1-week ahead forecast value of confirmed cases for 2024-04-28 is: 79 cases/week
The 95 % prediction interval for this forecast is (50-116)
The 50 % prediction interval for this forecast is (70-91)
The last estimated Rt value with the 95% confidence interval is: 1.03 (0.96-1.1)
The 1-week ahead forecast value of confirmed cases for 2024-04-28 is: 11 cases/week
The 95 % prediction interval for this forecast is (1-30)
The 50 % prediction interval for this forecast is (6-16)
The last estimated Rt value with the 95% confidence interval is: 0.88 (0.77-0.99)
The 1-week ahead forecast value of confirmed cases for 2024-04-28 is: 25 cases/week
The 95 % prediction interval for this forecast is (15-37)
The 50 % prediction interval for this forecast is (21-29)
The last estimated Rt value with the 95% confidence interval is: 0.78 (0.7-0.86)
The 1-week ahead forecast value of confirmed cases for 2024-05-05 is: 4 cases/week
The 95 % prediction interval for this forecast is (0-16)
The 50 % prediction interval for this forecast is (2-7)
The last estimated Rt value with the 95% confidence interval is: 0.78 (0.6-0.96)
The 1-week ahead forecast value of confirmed cases for 2024-05-05 is: 3 cases/week
The 95 % prediction interval for this forecast is (0-10)
The 50 % prediction interval for this forecast is (1-5)
The last estimated Rt value with the 95% confidence interval is: 0.62 (0.47-0.77)
The 1-week ahead forecast value of confirmed cases for 2024-05-05 is: 6 cases/week
The 95 % prediction interval for this forecast is (0-25)
The 50 % prediction interval for this forecast is (2-11)
The last estimated Rt value with the 95% confidence interval is: 1 (0.77-1.23)
The 1-week ahead forecast value of confirmed cases for 2024-05-05 is: 5 cases/week
The 95 % prediction interval for this forecast is (1-11)
The 50 % prediction interval for this forecast is (3-7)
The last estimated Rt value with the 95% confidence interval is: 0.31 (0.12-0.5)
This section presents validation plots for the 1-week ahead model forecasts for the current season for both PHRDW and PLOVER data. The forecasting model is trained by fitting to historical data, which is used to then make the prediction 1-week ahead. Here, the fit of the model to the known historical data can be visualized. Overlap between the prediction interval and the historical cases indicates the performance of the model to forecast new cases over a 1 week time horizon
This section presents validation metrics for the 1-week ahead model forecasts for the current season for both PHRDW and PLOVER data. The forecasting model is trained by fitting to historical data, which is used to then make the prediction 1-week ahead. Here, the fit of the model to the known historical data is summarized. For each of the weekly historical forecasts, the metrics output whether the forecasted confirmed cases lie in the 50 or 95% percentile interval as well as the mean-squared percentage error. A higher mean squared percentage error and a relatively larger proportion of forecasts outside the 95% percentile interval are indicative of poor quality historical (and thus likely current) forecasts.
Previous 1-week ahead forecasts had 38.97 % (53/136) of the true confirmed cases within the 50% prediction interval and 62.5 % (32/136) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 174.574
In the 2021 season, Previous 1-week ahead forecasts had 6.67 % (1/15) of the true confirmed cases within the 50% prediction interval and 33.34 % (4/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 108.269
In the 2022 season, Previous 1-week ahead forecasts had 19.23 % (10/52) of the true confirmed cases within the 50% prediction interval and 44.23 % (13/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 254.633
In the 2023 season, Previous 1-week ahead forecasts had 69.81 % (37/53) of the true confirmed cases within the 50% prediction interval and 86.79 % (9/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 96.746
Previous 1-week ahead forecasts had 11.76 % (16/136) of the true confirmed cases within the 50% prediction interval and 33.08 % (29/136) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 19.803
In the 2021 season, Previous 1-week ahead forecasts had 20 % (3/15) of the true confirmed cases within the 50% prediction interval and 26.67 % (1/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 32.418
In the 2022 season, Previous 1-week ahead forecasts had 1.92 % (1/52) of the true confirmed cases within the 50% prediction interval and 9.61 % (4/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 18.757
In the 2023 season, Previous 1-week ahead forecasts had 18.87 % (10/53) of the true confirmed cases within the 50% prediction interval and 58.49 % (21/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 8.485
Previous 1-week ahead forecasts had 17.65 % (24/136) of the true confirmed cases within the 50% prediction interval and 44.12 % (36/136) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 20.961
In the 2021 season, Previous 1-week ahead forecasts had 6.67 % (1/15) of the true confirmed cases within the 50% prediction interval and 13.34 % (1/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 41.119
In the 2022 season, Previous 1-week ahead forecasts had 17.31 % (9/52) of the true confirmed cases within the 50% prediction interval and 42.31 % (13/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 19.91
In the 2023 season, Previous 1-week ahead forecasts had 24.53 % (13/53) of the true confirmed cases within the 50% prediction interval and 56.61 % (17/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 13.815
Previous 1-week ahead forecasts had 62.5 % (85/136) of the true confirmed cases within the 50% prediction interval and 75.74 % (18/136) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 139.414
In the 2021 season, Previous 1-week ahead forecasts had 13.33 % (2/15) of the true confirmed cases within the 50% prediction interval and 86.67 % (13/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 319.938
In the 2022 season, Previous 1-week ahead forecasts had 84.62 % (44/52) of the true confirmed cases within the 50% prediction interval and 88.47 % (2/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 125.355
In the 2023 season, Previous 1-week ahead forecasts had 62.26 % (33/53) of the true confirmed cases within the 50% prediction interval and 84.9 % (12/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 54.73
Previous 1-week ahead forecasts had 29.2 % (40/137) of the true confirmed cases within the 50% prediction interval and 52.56 % (32/137) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 39.187
In the 2021 season, Previous 1-week ahead forecasts had 26.67 % (4/15) of the true confirmed cases within the 50% prediction interval and 40 % (2/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 74.674
In the 2022 season, Previous 1-week ahead forecasts had 17.31 % (9/52) of the true confirmed cases within the 50% prediction interval and 30.77 % (7/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 39.768
In the 2023 season, Previous 1-week ahead forecasts had 32.08 % (17/53) of the true confirmed cases within the 50% prediction interval and 73.59 % (22/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 24.603
Previous 1-week ahead forecasts had 1.46 % (2/137) of the true confirmed cases within the 50% prediction interval and 7.3 % (8/137) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 26.091
In the 2021 season, Previous 1-week ahead forecasts had 6.67 % (1/15) of the true confirmed cases within the 50% prediction interval and 33.34 % (4/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 53.43
In the 2022 season, Previous 1-week ahead forecasts had 100 % (52/52) of the true confirmed cases within the 50% prediction interval and NA % (NA/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 25.386
In the 2023 season, Previous 1-week ahead forecasts had 3.77 % (2/53) of the true confirmed cases within the 50% prediction interval and 96.23 % (51/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 16.588
Previous 1-week ahead forecasts had 14.6 % (20/137) of the true confirmed cases within the 50% prediction interval and 34.31 % (27/137) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 38.426
In the 2021 season, Previous 1-week ahead forecasts had 6.67 % (1/15) of the true confirmed cases within the 50% prediction interval and 20 % (2/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 65.23
In the 2022 season, Previous 1-week ahead forecasts had 15.38 % (8/52) of the true confirmed cases within the 50% prediction interval and 32.69 % (9/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 48.066
In the 2023 season, Previous 1-week ahead forecasts had 16.98 % (9/53) of the true confirmed cases within the 50% prediction interval and 32.07 % (8/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 18.094
Previous 1-week ahead forecasts had 51.09 % (70/137) of the true confirmed cases within the 50% prediction interval and 76.64 % (35/137) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 32.506
In the 2021 season, Previous 1-week ahead forecasts had 6.67 % (1/15) of the true confirmed cases within the 50% prediction interval and 13.34 % (1/15) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 85.905
In the 2022 season, Previous 1-week ahead forecasts had 78.85 % (41/52) of the true confirmed cases within the 50% prediction interval and 86.54 % (4/52) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 18.695
In the 2023 season, Previous 1-week ahead forecasts had 35.85 % (19/53) of the true confirmed cases within the 50% prediction interval and 81.13 % (24/53) of the true confirmed cases in the 95% prediction interval
The Mean squared percentage error on the validation set is: 15.566